Google / DeepMind

Gemini and related research lines across Google Brain and DeepMind. Gemini 等模型与相关研究线（Google Brain / DeepMind）。

Official site ↗

Models in the tree

25 featured nodes · 5 other nodes (from Google, Google DeepMind, DeepMind, Google Brain ). (54 collapsed variants hidden from the graph)

Featured (graph/timeline)

• Gemma 3n — Series — Apr 2026
• Gemini 3.1 — Series — Apr 2026
• Gemma 4 — Series — Apr 2026
• Lyria 3 — Google's Generative Music + Live Mode — Feb 2026
• PALM-2 — Jan 2026
• Gemini 3 Pro — Native Omni-Modal Frontier — Nov 2025
• Nano Banana Pro / Gemini 3 Pro Image — 4K Text-Native Image — Nov 2025
• Genie 3 — Real-Time Interactive World Model — Aug 2025
• Veo 3 — Google's Native-Audio Video Model — May 2025
• Gemini 2.5 Pro — Google's Native Reasoning Frontier — Mar 2025
• Gemini Robotics — VLA from DeepMind — Mar 2025
• Gemma 3 — Google's Open-Weight Family — Mar 2025
• AlphaProof — Math Olympiad Reasoning — Jul 2024
• AlphaFold 3 — All-Atom Biomolecular Prediction — May 2024
• Gemini 1 — Google's Native Multimodal Frontier — Dec 2023
• GraphCast — AI Weather Forecasting at ECMWF Quality — Nov 2023
• ReAct — Reasoning + Acting in Language Models — Oct 2022
• PaLM — Pathways Language Model (540B) — Apr 2022
• Chinchilla — Compute-Optimal LLM Training — Mar 2022
• Chain-of-Thought Prompting (CoT) — Jan 2022
• AlphaFold 2 — Protein Structure Prediction Solved — Jul 2021
• BERT — Bidirectional Encoder Representations from Transformers — Oct 2018
• Attention Is All You Need (Transformer) — Jun 2017
• AlphaGo Defeats Lee Sedol — Deep RL at the Game Frontier — Mar 2016
• …and 1 more

Series

Benchmark snapshot

Best score within this company’s nodes (not a normalized cross-benchmark ranking).

• Speed · Output tok/s: 122 (from gemini-3)
• GPQA: 89.8% (from gemini-3)
• MMLU-Pro: 89.0% (from gemini-3)
• Intelligence · AA Intelligence Index: 51.40 (from gemini-1-5)
• Coding · SciCode: 50.6% (from gemini-3)
• Coding · AA Coding Index: 43.40 (from gemini-1-5)
• Agentic · Terminal-Bench Hard: 43.2% (35.6–43.2%) (from gemini-1-5)
• HLE: 34.7% (from gemini-3)

Official catalog (manual)

• gemini-3-pro — Gemini 3 Pro — in tree
• gemini-3-pro-image — Gemini 3 Pro Image — in tree
• gemini-3-flash — Gemini 3 Flash — in tree
• gemini-3.1-pro — Gemini 3.1 Pro (Preview) — in tree
• gemini-3.1-flash-live — Gemini 3.1 Flash Live — in tree
• gemini-3.1-flash-lite — Gemini 3.1 Flash-Lite — in tree
• gemini-3.1-flash-image — Gemini 3.1 Flash Image — in tree
• gemini-3.1-flash-audio — Gemini 3.1 Flash Audio — in tree
• gemma-4-26b-a4b — Gemma 4 26B A4B (Reasoning) — in tree
• gemma-4-31b — Gemma 4 31B (Reasoning) — in tree
• gemini-3-1-pro-preview — Gemini 3.1 Pro Preview — in tree
• gemma-4-31b-non-reasoning — Gemma 4 31B (Non-reasoning) — in tree
• gemini-2-5-pro — Gemini 2.5 Pro — in tree
• gemma-4-e4b-non-reasoning — Gemma 4 E4B (Non-reasoning) — in tree
• gemma-4-e2b — Gemma 4 E2B (Reasoning) — in tree
• gemma-4-e4b — Gemma 4 E4B (Reasoning) — in tree
• gemini-3-deep-think — Gemini 3 Deep Think — in tree
• gemini-3-flash-reasoning — Gemini 3 Flash Preview (Reasoning) — in tree
• gemini-2-5-flash-lite-preview-09-2025-reasoning — Gemini 2.5 Flash-Lite Preview (Sep '25) (Reasoning) — in tree
• gemini-2-5-flash-lite-preview-09-2025 — Gemini 2.5 Flash-Lite Preview (Sep '25) (Non-reasoning) — in tree
• gemini-3-pro-low — Gemini 3 Pro Preview (low) — in tree
• gemma-4-e2b-non-reasoning — Gemma 4 E2B (Non-reasoning) — in tree
• gemma-4-26b-a4b-non-reasoning — Gemma 4 26B A4B (Non-reasoning) — in tree
• gemma-3-270m — Gemma 3 270M — in tree
• gemini-3-1-flash-lite-preview — Gemini 3.1 Flash-Lite Preview — in tree
• gemini-2-0-pro-experimental-02-05 — Gemini 2.0 Pro Experimental (Feb '25) — in tree
• gemini-2-0-flash-experimental — Gemini 2.0 Flash (experimental) — in tree
• gemini-1-5-pro — Gemini 1.5 Pro (Sep '24) — in tree
• gemini-2-0-flash-lite-preview — Gemini 2.0 Flash-Lite (Preview) — in tree
• gemini-2-0-flash — Gemini 2.0 Flash (Feb '25) — in tree
• gemini-1-5-flash — Gemini 1.5 Flash (Sep '24) — in tree
• gemini-1-5-flash-8b — Gemini 1.5 Flash-8B — in tree
• gemini-2-5-flash — Gemini 2.5 Flash (Non-reasoning) — in tree
• gemini-1-5-pro-may-2024 — Gemini 1.5 Pro (May '24) — in tree
• gemma-3-1b — Gemma 3 1B Instruct — in tree
• gemma-3n-e2b — Gemma 3n E2B Instruct — in tree
• gemma-3-12b — Gemma 3 12B Instruct — in tree
• gemini-1-5-flash-may-2024 — Gemini 1.5 Flash (May '24) — in tree
• gemini-2-5-flash-reasoning — Gemini 2.5 Flash (Reasoning) — in tree
• gemma-3-27b — Gemma 3 27B Instruct — in tree
• gemma-3-4b — Gemma 3 4B Instruct — in tree
• gemini-2-5-flash-preview-09-2025-reasoning — Gemini 2.5 Flash Preview (Sep '25) (Reasoning) — in tree
• gemini-2-5-flash-lite — Gemini 2.5 Flash-Lite (Non-reasoning) — in tree
• gemini-2-5-flash-preview-09-2025 — Gemini 2.5 Flash Preview (Sep '25) (Non-reasoning) — in tree
• gemini-1-0-pro — Gemini 1.0 Pro — in tree
• gemma-3n-e4b — Gemma 3n E4B Instruct — in tree
• gemini-2-5-flash-reasoning-04-2025 — Gemini 2.5 Flash Preview (Reasoning) — in tree
• gemini-2-5-pro-03-25 — Gemini 2.5 Pro Preview (Mar' 25) — in tree
• gemini-2-0-flash-thinking-exp-1219 — Gemini 2.0 Flash Thinking Experimental (Dec '24) — in tree
• gemini-2-0-flash-lite-001 — Gemini 2.0 Flash-Lite (Feb '25) — in tree
• gemini-2-5-flash-04-2025 — Gemini 2.5 Flash Preview (Non-reasoning) — in tree
• gemini-2-5-pro-05-06 — Gemini 2.5 Pro Preview (May' 25) — in tree
• gemini-2-0-flash-thinking-exp-0121 — Gemini 2.0 Flash Thinking Experimental (Jan '25) — in tree
• gemma-3n-e4b-preview-0520 — Gemma 3n E4B Instruct Preview (May '25) — in tree
• gemini-1-0-ultra — Gemini 1.0 Ultra — in tree
• gemini-2-5-flash-lite-reasoning — Gemini 2.5 Flash-Lite (Reasoning) — in tree
• palm-2 — PALM-2 — in tree

Sources

• https://deepmind.google/
• https://deepmind.google/models/model-cards
• https://blog.google/products/gemini/gemini-3/
• https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/

Notes

This page groups the historical split between Google Brain and DeepMind under a single company view.